Predicting the Bumper Crops of the Future

Richer harvest. A new model of crop yields will help predict the possible size of harvests, and help provide advice to farmers such as this woman in Indonesia.

Wikimedia Commons

BOSTON—Farms in the United States are a miracle of productivity and the source of enormous exports of grain, but they can only do so much to feed the world. In many parts of the globe, crops yields are just a fraction of what would be possible if they had more fertilizer, better seeds, and improved management. A $2 million project called the Global Yield Gap Atlas, described in a session here today at the annual meeting of AAAS (which publishes ScienceNOW), is attempting to figure out exactly how much food could be grown on the planet. It's a novel effort involving agronomists from around the world.

"Efforts like this are badly needed," writes Jon Foley, a global change scientist at University of Minnesota, Twin Cities, in an e-mail. "We know that the pressures of increasing food security will only get bigger as population and economic growth pressure the world's food system," writes Foley, who is not involved in the work, "and understanding how and where yields can still improve is vital to improving food security."

Other experts say the results will ultimately be useful to regional farm managers, national policymakers, and to food security experts studying global trends. In the session, one of the leaders of the atlas described recent results that suggest the ambitious goals are within reach.

This isn't the first attempt to model the shortcomings of current agricultural production, called "yield gaps." International institutions such as CGIAR have developed computer models that use data on climate, crop types, and other factors to estimate current food production and forecast future trends. But the new project aims to improve these estimates. "Previous efforts are not accurate or robust enough," said Kenneth Cassman, an agronomist at the University of Nebraska, Lincoln, and a co-leader of the atlas.

The way to improve is to take a bottom-up approach by first gathering detailed local data on weather, soil, cultivars, and farming practices. Then these data will be plugged into a computer model that will extrapolate the data to broader areas based on climatic data. Work began in 2011, funded primarily by a grant from the Bill & Melinda Gates Foundation. The group is now getting data from agronomists in South America, South Asia, and sub-Saharan Africa.

At the meeting, Cassman described initial results. The team has found that it can predict yields for 80% of the acres planted with corn in the United States by knowing just five climatic zones. And it can get a robust estimate of rain-fed corn around the world by using 49 climate zones. To Cassman, this means that predicting yields and calculating yield gaps is "a tractable problem." The results are described in three recent papers, published in Field Crops Research. The goal is to have the atlas completed within 4 years.

"The project is an excellent effort," according to sustainability scientist Nathan Mueller of the University of Minnesota, Twin Cities. "It is really the next step that's needed." Montague Demment, who studies international agricultural development at the Association of Public and Land-grant Universities in Washington, D.C., calls the atlas "absolutely unique" and predicts that it will help set research priorities and bring attention to the need to invest in raising yields.

Follow News from Science